etc/yt-dlp
0
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-12-19 05:29:59 +00:00
Code Issues Projects Releases Wiki Activity

[ie/bilibili] Fix extractor (#11667)

Browse source 
Closes #11665
Authored by: grqz
This commit is contained in:
N/Ame 2024-12-02 10:55:18 +13:00 committed by GitHub
parent 0d146c1e36
commit 239f5f36fe
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 21 additions and 22 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

39
yt_dlp/extractor/bilibili.py
View file

@ -18,7 +18,6 @@
InAdvancePagedList, InAdvancePagedList,
OnDemandPagedList, OnDemandPagedList,
bool_or_none, bool_or_none,
clean_html,
determine_ext, determine_ext,
filter_dict, filter_dict,
float_or_none, float_or_none,
@ -639,31 +638,27 @@ def _real_extract(self, url):
headers['Referer'] = url headers['Referer'] = url
initial_state = self._search_json(r'window\.__INITIAL_STATE__\s*=', webpage, 'initial state', video_id) initial_state = self._search_json(r'window\.__INITIAL_STATE__\s*=', webpage, 'initial state', video_id)
is_festival = 'videoData' not in initial_state
if is_festival:
video_data = initial_state['videoInfo']
else:
play_info_obj = self._search_json(
r'window\.__playinfo__\s*=', webpage, 'play info', video_id, fatal=False)
if not play_info_obj:
if traverse_obj(initial_state, ('error', 'trueCode')) == -403: if traverse_obj(initial_state, ('error', 'trueCode')) == -403:
self.raise_login_required() self.raise_login_required()
if traverse_obj(initial_state, ('error', 'trueCode')) == -404: if traverse_obj(initial_state, ('error', 'trueCode')) == -404:
raise ExtractorError( raise ExtractorError(
'This video may be deleted or geo-restricted. ' 'This video may be deleted or geo-restricted. '
'You might want to try a VPN or a proxy server (with --proxy)', expected=True) 'You might want to try a VPN or a proxy server (with --proxy)', expected=True)
play_info = traverse_obj(play_info_obj, ('data', {dict}))
if not play_info: is_festival = 'videoData' not in initial_state
if traverse_obj(play_info_obj, 'code') == 87007: if is_festival:
toast = get_element_by_class('tips-toast', webpage) or '' video_data = initial_state['videoInfo']
msg = clean_html( else:
f'{get_element_by_class("belongs-to", toast) or ""}，'
+ (get_element_by_class('level', toast) or ''))
raise ExtractorError(
f'This is a supporter-only video: {msg}. {self._login_hint()}', expected=True)
raise ExtractorError('Failed to extract play info')
video_data = initial_state['videoData'] video_data = initial_state['videoData']
if video_data.get('is_upower_exclusive'):
high_level = traverse_obj(initial_state, ('elecFullInfo', 'show_info', 'high_level', {dict})) or {}
raise ExtractorError(
'This is a supporter-only video: '
f'{join_nonempty("title", "sub_title", from_dict=high_level, delim="，")}. '
f'{self._login_hint()}', expected=True)
video_id, title = video_data['bvid'], video_data.get('title') video_id, title = video_data['bvid'], video_data.get('title')
# Bilibili anthologies are similar to playlists but all videos share the same video ID as the anthology itself. # Bilibili anthologies are similar to playlists but all videos share the same video ID as the anthology itself.
@ -689,10 +684,14 @@ def _real_extract(self, url):
old_video_id = format_field(aid, None, f'%s_part{part_id or 1}') old_video_id = format_field(aid, None, f'%s_part{part_id or 1}')
cid = traverse_obj(video_data, ('pages', part_id - 1, 'cid')) if part_id else video_data.get('cid') cid = traverse_obj(video_data, ('pages', part_id - 1, 'cid')) if part_id else video_data.get('cid')
play_info = (
traverse_obj(
self._search_json(r'window\.__playinfo__\s*=', webpage, 'play info', video_id, default=None),
('data', {dict}))
or self._download_playinfo(video_id, cid, headers=headers))
festival_info = {} festival_info = {}
if is_festival: if is_festival:
play_info = self._download_playinfo(video_id, cid, headers=headers)
festival_info = traverse_obj(initial_state, { festival_info = traverse_obj(initial_state, {
'uploader': ('videoInfo', 'upName'), 'uploader': ('videoInfo', 'upName'),
'uploader_id': ('videoInfo', 'upMid', {str_or_none}), 'uploader_id': ('videoInfo', 'upMid', {str_or_none}),